Can subscript be extracted

[fsmosca]

I have a pdf containing a table with subscript text. I tried to extract it, converted it to csv using pandas and save it with utf-8-sig encoding. But it is not working. There is a newline char inserted.

import pdfplumber
import pandas as pd

pdf = pdfplumber.open("subscript.pdf")
table = pdf.pages[0].extract_table()
df = pd.DataFrame(table)
df.to_csv('table.csv', index=False, encoding='utf-8-sig')
print(df)

output

          0
0       aaa
1     NH\n4
2  C\n10-13

How to eliminate the newline char \n in table creation?

This should have been NH4 and C10-13.

Note if the subscript in the pdf is really a unicode symbol, then the library can handle it.

[samkit-jain]

Hi @fsmosca Appreciate your interest in the library. You'll have to write your own logic to identify what characters are subscript or superscript based on their size attribute

Closing as duplicate of #730

[fsmosca]

I recommend examining the size and bottom/top attributes of the objects in Page.chars.

Is there a sample guide to do that?

[samkit-jain]

I would expect the subscript chars to be smallest in size. You can use the following code

import pdfplumber

pdf = pdfplumber.open("file.pdf")
p = pdf.pages[0]

# Smallest character size.
min_size = min(c["size"] for c in p.chars)

# Draw those characters.
im = p.to_image(resolution=200)
im.draw_rects([char for char in p.objects['char'] if char['size'] == min_size])
im.save("image.png", format="PNG")